Back to Table of contents

Primeur weekly 2012-11-19

Exascale supercomputing

DataDirect Networks announces $100 million investment in Exascale computing ...

Intel and DataDirect Networks join forces to enable compute infrastructures 1,000 times larger than today's biggest systems ...

The Cloud

HP delivers industry's first purpose-built server for Big Data ...

DDN adds enterprise class Cloud Tiering, NAS and SSD capabilities to its parallel file system appliances, making them Big Data ready ...

Desktop Grids

Members of the public asked to help tend Feynman's Flowers ...

EuroFlash

EMI 2 - Matterhorn - 5th update is now available ...

German Federal Waterways Engineering and Research Institute (BAW) optimizes navigable waterways using a Bull supercomputer ...

Choreographing light ...

USFlash

New editions of Moab HPC Suite enhance HPC system efficiency ...

SGI ICE X selected by CRIEPI for electric power research ...

Supermicro debuts SuperServer and SuperBlade solutions supporting new AMD Opteron 6300 Series processor ...

Brocade collaborates with Dell to announce industry's first 16 Gbps Fibre Channel end-to-end solution ...

Department of Energy's ESnet rolls out world's fastest science network ...

New storage and management features for IBM PureSystems help clients reduce costs and increase utilization ...

Obsidian welcomes Mellanox and its MetroX product into the long haul InfiniBand ...

Commercial and technical HPC customers can achieve Big Data results faster with NetApp HPS Rack solution ...

Brocade and Hyve Solutions partner to deliver fully integrated Big Data solutions using industry-leading Brocade VCS Fabric network technology ...

AMD FirePro server graphics redefining supercomputing performance through power efficient, sustainable technology ...

Medic Vision introduces industry's first enterprise-wide iterative image reconstruction system ...

StackIQ announces release of StackIQ Enterprise HPC for high performance computing clusters ...

Heterogeneous systems re-claim Green500 list dominance ...

PNNL expertise highlighted at Supercomputing ...

Keeneland Project deploys new GPU supercomputing system for the National Science Foundation ...

Titan is also a green powerhouse ...

AMD delivers massive compute performance for world's top ranked supercomputer ...

Department of Energy's ESnet rolls out world's fastest science network

13 Nov 2012 Berkeley - As scientific research becomes increasingly data-intensive and globally collaborative, the ability to quickly and reliably share enormous data sets is critical to challenges as diverse as finding cleaner energy technologies, understanding climate change and investigating the nature of our universe. To accelerate such discovery, the U.S. Department of Energy's (DOE) ESnet (Energy Sciences Network) is now operating the world's fastest science network, serving the entire national laboratory system, its supercomputing centres, and its major scientific instruments at 100 gigabits per second - 10 times faster than its previous generation network. ESnet is managed by Lawrence Berkeley National Laboratory (Berkeley Lab).

"In January, 1983, the Internet was launched into operation on a 50 Kb/s ARPANET backbone. 30 years later, ESnet backbone speeds are two million times faster", stated ESnet Policy Board member Vint Cerf, Google's chief Internet evangelist and Internet pioneer. "Terabit speeds are in sight and faster aggregate speeds can be anticipated. Talk about riding a rocket."

"The upgrade, funded by DOE's Office of Science, comes just in time", stated ESnet Director Greg Bell, who also leads Berkeley Lab's Scientific Networking Division. "ESnet traffic doubles every 18 months, roughly twice as fast as the commercial Internet. Based on analysis of its traffic over the years and future plans, the amount of data carried by ESnet will reach over 100 petabytes per month by 2016. That's because we're moving into a new era of extreme data, and scientific discovery in many fields depends on data mobility. ESnet is becoming something more than a network infrastructure - it's now an extension of the experimental facilities it serves, and a vital component in the process of discovery."

"Like many other scientific disciplines, cosmology is increasingly reliant on collecting, sharing and analyzing massive datasets, whether from instruments launched into space, from ground-based telescopes or detailed simulations", stated Berkeley Lab cosmologist Peter Nugent and member of the team that won the 2011 Nobel Prize in Physics. "Advanced networking, much like advanced computing, is a critical component for the broad collaborations that are behind many discoveries. ESnet's move to 100G connectivity is especially important as projects like the Baryon Oscillation Spectroscopic Survey and Dark Energy Survey generate unprecedented amounts of data, which is stored at Berkeley Lab and then served up to the global research community."

ESnet partnered with Internet2, the advanced networking and technology community that connects America's universities and research institutions and extends connectivity to research networks worldwide, to deploy its 100 Gbps network over a new, highly-scalable optical infrastructure that the two organisations share for the benefit of their respective communities.

"Big Data science and collaborative research are the current and future reality", stated Internet2 President and CEO Dave Lambert. "Through this key partnership, 100G technologies are extended throughout the national research infrastructure - creating a much richer and more robust national and global scientific collaboration environment - from which new innovations will be born, great challenges facing all of society will be confronted, and discovery will be accelerated."

To pave the way for the production network, last year ESnet announced the completion of a 100 Gbps prototype network - funded by the American Recovery and Reinvestment Act - which connected the three leading DOE Office of Science supercomputing sites as a proof of concept. With the 100 Gbps national network now in full production across its entire footprint, the ESnet team is moving to upgrade individual lab connections to enable them to take even greater advantage of the speed and capacity of the backbone.

"100 Gbps is just the beginning", Greg Bell stated. "Because we can scale far beyond that capacity on the underlying optical platform we have invested in with Internet2, ESnet will continue to serve DOE science for years to come."
Source: DOE/Lawrence Berkeley National Laboratory

Back to Table of contents

Primeur weekly 2012-11-19

Exascale supercomputing

DataDirect Networks announces $100 million investment in Exascale computing ...

Intel and DataDirect Networks join forces to enable compute infrastructures 1,000 times larger than today's biggest systems ...

The Cloud

HP delivers industry's first purpose-built server for Big Data ...

DDN adds enterprise class Cloud Tiering, NAS and SSD capabilities to its parallel file system appliances, making them Big Data ready ...

Desktop Grids

Members of the public asked to help tend Feynman's Flowers ...

EuroFlash

EMI 2 - Matterhorn - 5th update is now available ...

German Federal Waterways Engineering and Research Institute (BAW) optimizes navigable waterways using a Bull supercomputer ...

Choreographing light ...

USFlash

New editions of Moab HPC Suite enhance HPC system efficiency ...

SGI ICE X selected by CRIEPI for electric power research ...

Supermicro debuts SuperServer and SuperBlade solutions supporting new AMD Opteron 6300 Series processor ...

Brocade collaborates with Dell to announce industry's first 16 Gbps Fibre Channel end-to-end solution ...

Department of Energy's ESnet rolls out world's fastest science network ...

New storage and management features for IBM PureSystems help clients reduce costs and increase utilization ...

Obsidian welcomes Mellanox and its MetroX product into the long haul InfiniBand ...

Commercial and technical HPC customers can achieve Big Data results faster with NetApp HPS Rack solution ...

Brocade and Hyve Solutions partner to deliver fully integrated Big Data solutions using industry-leading Brocade VCS Fabric network technology ...

AMD FirePro server graphics redefining supercomputing performance through power efficient, sustainable technology ...

Medic Vision introduces industry's first enterprise-wide iterative image reconstruction system ...

StackIQ announces release of StackIQ Enterprise HPC for high performance computing clusters ...

Heterogeneous systems re-claim Green500 list dominance ...

PNNL expertise highlighted at Supercomputing ...

Keeneland Project deploys new GPU supercomputing system for the National Science Foundation ...

Titan is also a green powerhouse ...

AMD delivers massive compute performance for world's top ranked supercomputer ...